applicable to the Gold Bullion Assay. 101 



parison of two ponderable bodies is effected in such a manner 

 as to eliminate the errors proper to the balance itself. 



In Borda's method : — The weights to be compared may 

 be distinguished as Nos. 1 and 2. We counterpoise the 

 heavier of the two weights by shot in the opposite pan ; we 

 now substitute for weight No. 1 that marked No. 2, and add 

 to the latter small fractional weights until an equipoise is 

 established. The weights added show how much No. 1 is 

 heavier than No. 2, and the result thus obtained is quite 

 independent of inequality of the arms of the beam. 



In Gauss' method : — The weights are placed in the 

 opposite pans and the difference noted ; the pans with 

 weights are now reversed on the beam, and the difference 

 again noted ; the mean of these differences is the real 

 difference of the weights in air, independent of any 

 inequality of the arms of the balance. 



These examples of methods actually employed whenever 

 rigid determinations of weight are the object, show, beyond 

 question, that with balances rendered sensitive by lightness 

 and judicious distribution of metal, and by sharpness and 

 true position of the knife-edges, the accuracy of our work 

 will depend less on the accuracy of construction of the beam 

 itself than on the accuracy of the weights, and that a 

 method involving the principle of double weighing and 

 depending especially on the accuracy of the weights is the 

 one at present to be sought for as likely to lead to a nearer 

 and more uniform approach to accuracy of results. 



In proposing a mode of assay which involves the 

 principles of double weighing, I wish to explain that 

 although the method has been the subject of actual experi- 

 ment, it is yet quite new in my hands ; I venture to 

 describe it in the sense of a proposition, the adoption of 

 which into daily practice mast depend upon its ultimate 

 proved merits as compared with the routine of weighing at 

 present in use. 



I will now make a short explanation concerning the 

 ■assay er's weights, those employed at the present time. Let 

 us take the case in which the assay is made on 10 grains of 

 each sample, the weight representing this amount taken for 

 trial, is what is commonly called the " assay pound." What- 

 ever proportion of pure gold this assay pound of the sample 

 is found to contain, such a proportion will the original bar, 

 from which the assay piece has been cut, contain : all 

 results being expressed according to a decimal notation. 



